A Fast Contingency Screening Technique for Generation System Reliability Evaluation

In most existing reliability evaluation techniques, the stopping rules applied for contingency enumeration are either a given number of states with the highest probabilities or the states up to a specified failure order. However, the probability of a high order outage state may be larger than that of a low order outage state due to different reliabilities of the failed components. On the other hand, a low probability state with high outage capacity may have larger impact on system reliability than a high probability state with low outage capacity. These effects have not been considered in the existing techniques for contingency filtering or screening. A fast contingency screening technique (FCST) for generation system reliability evaluation is proposed in this paper. In this technique, contingency states are searched and ranked based on the severity coefficients determined by both state probabilities and outage capacities of the failed units. Several severity coefficients for a unit are defined and examined based on reliability and capacity factors. The replaceable neighboring states of a state and the minimum set of the neighboring states are defined to further reduce computation time. The merits of the proposed techniques and severity coefficients are validated by evaluating the East-China Power Grid.